On Mon, 16 Aug 2004 13:57:13 -0400, "Michael Stassen" <[EMAIL PROTECTED]> said: > You were perfectly clear. We understand that you only want to test f2 > and f3 for uniqueness. The question is, which of the possible values > of f1 do you want to get. Do you see? For a particular unique f2, f3 > combination, there may be multiple f1 values. How should we choose > which one to put in the new table?
Oh, I understand now, sorry. If I said "it makes no difference" then you'd ask what the heck I have f1 for in the first place...It actually doesn't make a difference. Maybe I should drop f1. f1 is an auto-increment int. so I imagine I'd want f1 re-incremented in numerical order to take the gaps out. Not exactly normalized (or normal:^), thanks. That is what Shawn has asked > twice, and you have not answered. Until you answer that, no one can > provide a correct solution. > > Michael > > leegold wrote: > > > On Mon, 16 Aug 2004 12:39:32 -0400, [EMAIL PROTECTED] said: > > > >>Let me see if I can explain it a little better....If you need to > >>move all 3 columns to the new table but you only want *1* row where > >>f2 and f3 have a unique combination of values, how do you want to > >>choose *which* value of f1 to move over with that combination? Do > >>you want the minimum value, the maximum value, or no value at all? > > > > > > Whoa, it's not that complicated....I want to test only f2 && f3 for > > uniqueness, not f1 && f2 && f3. That's all. If I'm not making it > > clear - don't worry...it's not life or death. Thanks. ...snip... > > > -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED]